Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Build an evidence-backed ECC install plan for a specific repo by sorting skills, commands, rules, hooks, and extras into DAILY vs LIBRARY buckets using parallel repo-aware review passes. Use when ECC should be trimmed to what a project actually needs instead of loading the full bundle.
.claude/skills/affaan-m-agent-sort/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-16 | ✗→✓ | ▲ Improved | 19% | 0% |
| case-17 | ✗→✓ | ▲ Improved | 73% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 42% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 31% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 80% | 0% |
当仓库需要项目特定的 ECC 表面而非默认完整安装时，使用此技能。
目标不是猜测"什么感觉有用"。目标是根据实际代码库中的证据对 ECC 组件进行分类。
按顺序生成以下工件：
仅使用两个分类：
DAILYLIBRARY在进行任何分类之前，使用仓库本地证据：
有用的命令包括：
bashrg --files rg -n "typescript|react|next|supabase|django|spring|flutter|swift" cat package.json cat pyproject.toml cat Cargo.toml cat pubspec.yaml cat go.mod
如果并行子代理可用，将审查分为以下轮次：
agents/*skills/*commands/*rules/*如果子代理不可用，则按顺序运行相同的轮次。
在分类任何内容之前，确定实际技术栈：
对于每个候选表面，记录：
使用此格式：
textskills/frontend-patterns | skill | DAILY | 84 个 .tsx 文件，存在 next.config.ts | 核心前端技术栈 skills/django-patterns | skill | LIBRARY | 无 .py 文件，无 pyproject.toml | 此仓库中未激活 rules/typescript/* | rules | DAILY | 存在 package.json + tsconfig.json | 活跃的 TS 仓库 rules/python/* | rules | LIBRARY | 零个 Python 源文件 | 仅保持可访问
提升至 DAILY 当：
降级至 LIBRARY 当：
将分类转化为行动：
.claude/skills/skill-library 保持可访问如果仓库已使用选择性安装，则更新该计划而非创建另一个系统。
如果项目需要可搜索的库表面，创建：
.claude/skills/skill-library/SKILL.md该路由器应包含：
不要在路由器内重复每个技能的主体。
应用计划后，验证：
返回一个简洁的报告，包含：
如果下一步是交互式安装或修复，交接至：
configure-ecc如果下一步是重叠清理或目录审查，交接至：
skill-stocktake如果下一步是更广泛的上下文修剪，交接至：
strategic-compact按此顺序返回结果：
text栈 - 语言/框架/运行时摘要 日常 - 始终加载的条目及证据 库 - 可搜索/参考的条目及证据 安装计划 - 应安装、移除或路由的内容 验证 - 已运行的检查及剩余差距
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 23,643 | 5,505 | -77% | 1 | 1 | 0% | 3,495 | 1,789 | -49% | 0 | 0 | — |
case-16 | fail→pass | 9,462 | 3,958 | -58% | 1 | 1 | 0% | 1,650 | 1,967 | +19% | 0 | 0 | — |
case-17 | fail→pass | 11,767 | 11,439 | -3% | 1 | 1 | 0% | 1,784 | 3,086 | +73% | 0 | 0 | — |
case-02 | fail→fail | 3,648 | 5,920 | +62% | 1 | 1 | 0% | 204 | 1,673 | +720% | 0 | 0 | — |
case-03 | fail→pass | 19,819 | 19,029 | -4% | 1 | 1 | 0% | 3,358 | 4,754 | +42% | 0 | 0 | — |
case-04 | fail→pass | 7,785 | 3,247 | -58% | 1 | 1 | 0% | 1,363 | 1,785 | +31% | 0 | 0 | — |
case-05 | fail→fail | 4,502 | 3,779 | -16% | 1 | 1 | 0% | 224 | 1,954 | +772% | 0 | 0 | — |
case-06 | fail→pass | 9,443 | 8,336 | -12% | 1 | 1 | 0% | 1,447 | 2,605 | +80% | 0 | 0 | — |
case-07 | fail→fail | 9,951 | 8,499 | -15% | 1 | 1 | 0% | 1,166 | 1,735 | +49% | 0 | 0 | — |
case-08 | fail→fail | 3,945 | 7,223 | +83% | 1 | 1 | 0% | 448 | 1,616 | +261% | 0 | 0 | — |
case-09 | fail→pass | 23,298 | 2,923 | -87% | 1 | 1 | 0% | 1,019 | 1,897 | +86% | 0 | 0 | — |
case-10 | pass→pass | 10,789 | 6,553 | -39% | 1 | 1 | 0% | 1,851 | 2,416 | +31% | 0 | 0 | — |
case-11 | fail→pass | 7,056 | 5,160 | -27% | 1 | 1 | 0% | 1,150 | 2,201 | +91% | 0 | 0 | — |
case-12 | pass→pass | 11,063 | 7,702 | -30% | 1 | 1 | 0% | 1,753 | 2,670 | +52% | 0 | 0 | — |
case-13 | pass→pass | 13,300 | 6,011 | -55% | 1 | 1 | 0% | 1,940 | 2,291 | +18% | 0 | 0 | — |
case-14 | fail→fail | 5,742 | 3,027 | -47% | 1 | 1 | 0% | 911 | 1,836 | +102% | 0 | 0 | — |
case-15 | pass→pass | 13,500 | 7,010 | -48% | 1 | 1 | 0% | 2,468 | 2,608 | +6% | 0 | 0 | — |
case-18 | fail→pass | 13,863 | 1,655 | -88% | 1 | 1 | 0% | 2,250 | 1,538 | -32% | 0 | 0 | — |
case-19 | fail→pass | 7,117 | 1,924 | -73% | 1 | 1 | 0% | 960 | 1,631 | +70% | 0 | 0 | — |
case-20 | pass→pass | 9,320 | 5,096 | -45% | 1 | 1 | 0% | 1,499 | 2,133 | +42% | 0 | 0 | — |
case-21 | fail→pass | 5,896 | 4,783 | -19% | 1 | 1 | 0% | 1,080 | 2,105 | +95% | 0 | 0 | — |
case-22 | fail→pass | 8,311 | 3,140 | -62% | 1 | 1 | 0% | 1,210 | 1,854 | +53% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted, and 17 counted toward the lift figure. The other 5 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +50 percentage points is the difference between those two pass rates over the 17 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
| Model | Method | Date | Lift |
|---|---|---|---|
| gemini-3.6-flash | verified | 8/3/2026 | +32% |
Other measured skills in the registry, with their headline benchmark lift.